1

2
nd
Int. PhD Symposium in Civil Engineering 1998 Budapest
HIGH STRENGTH LIGHTWEIGHT-AGGREGATE CONCRETE
Dipl.-Ing. Thorsten Faust
1
, Prof. Dr.-Ing. Dr.-Ing. e.h. Gert Knig
2
University of Leipzig, Institute for Structural Engineering and Building Material
D-04109 Leipzig
SUMMARY
One research field in Leipzig deals with the design of high strength lightweight-aggregate
concretes (HSLWAC). Due to the huge waveband of this concretes with all possible
combinations of lightweight-aggregates and mortar matrix the formulation of general
rules is difficult. Therefore both components lightweight-aggregates and mortar matrix
will be investigated in detail in order to achieve a deeper understanding of the internal
stress transfer and the failure mechanisms. Moreover the results of the components tests
serve as input data for a computer simulation. This instrument should support the
prediction and optimization of the compressive strength of LWAC.
Keywords: high strength lightweight-aggregate concrete, lightweight-aggregates, mortar
matrix, internal stress transfer, brittleness
1 INTRODUCTION
Lightweight-aggregate concrete with
closed structure is often called structural
LWAC with regard to the applications in
buildings, bridges and offshore structures.
The development of new concreting
materials in the last years expanded the
waveband of these concretes with the aim
to increase the strength or to decrease the
density, respectively. Fig 1 shows the huge
spectrum considering the strengths of 15 to
100 MPa and oven dry densities of 1,0 to
2,0 kg/dm. Structural LWAC with a
minimized density at a definite strength level is called HSLWAC. Thus the term high
strength in case of LWAC is not related to the strength, but to the relation of strength to
density. For all HSLWACs high strength mortar matrix are used so that in general the
concrete strength will be limited by the efficiency of the aggregates. This fact induces a
very brittle behaviour, which is independent of the strength. Consequently, the validity of
the existing design rules and calculation models have to be proven with regard to the use
of HSLWAC in order to adjust possible limits of applications.

1
PhD student
2
Professor in Structural Engineering
0
10
20
30
40
50
60
70
80
90
100
1 1,1 1,2 1,3 1,4 1,5 1,6 1,7 1,8 1,9 2
Oven-dry density [kg/dm]
Spectrum of structural lightweight aggregate concrete
C
u
b
e

c
o
m
p
r
e
s
s
i
v
e

s
t
r
e
n
g
t
h

[
N
/
m
m

]
Fig.1 Spectrum of structural
lightweight-aggregate concrete
2
2 LIGHTWEIGHT-AGGREGATE CONCRETES IN COMPRESSION
The stress/strain behaviour of LWAC in compression is, compared to normal density
concretes (NC) of the same compressive strength, generally characterized by a linear
ascending branch, a lower E-modulus and less ductility in the post-failure region. These
characteristics are usually more pronounced with increasing compressive strength and
decreasing oven dry density. Therefore HSLWAC not only comprises LWAC with a high
strength, but also LWAC with a low density.
Fig.2 shows a stress-strain
curve of a strain controlled
compressive test with a
cylindrical specimen. After
longitudinal cracking
(Fig.3a) pieces of concrete
spall off the cylinder
(Fig.3b). This phenomenon
was particularly observed in
case of ALWA-concretes
(All Light Weight
Aggregates). The reduced
cross sectional area requires a
quick unloading to avoid a
sudden failure. The
decreasing load allows an elastic expansion of the undestroyed regions outside the
damage zone. In case of brittle concretes combined with a small damage zone in relation
to the regions outside this fact causes a so called snap-back. In consideration of the low
modulus of elasticity this effect is valid for all HSLWAC.
Due to the snap-back effect the failure phenomena
of the specimen must be controlled by other suitable
means. The first experiments were performed with a
combination of axial and circumferential strain
control. This test setup was already used
successfully with HSC. But in case of LWAC,
partly unstable softening occurred as well, which
probably was caused by the low modulus of
elasticity (Fig.4). For this reason a control method,
where a considerable portion of the E-modulus will
be compensated, had to be taken into consideration.
This seems to be a promising solution in case of
LWAC [1,2].
Fig.5 shows the principle of the control method. A linear combination of deformation and
force is chosen as the feedback signal for a closed-loop servo-controlled system:
K
F
F
S
= .
Geometrically this control corresponds to a rotation of the axis with the angle =1/K,
which allows a steadily increasing feedback signal, if the stiffness value K is chosen
0
50
100
150
200
250
300
350
400
0,0 0,5 1,0 1,5 2,0 2,5 3,0 3,5 4,0 4,5 5,0
Strain[mm/m]
C
o
m
p
r
e
s
s
i
v
e

F
o
r
c
e

[
k
N
]
LC 45/55 =1,7 kg/dm
cylinder h/=300/100
circumferential expansion a
b
Fig. 2 Stress-strain curve Fig.3 Different fracture states
using strain control a) longitudinal cracks
b) spalling of concrete pieces
Fig. 4 Stress-strain curve using
axial and circumferential strain
control
3
between K
pre
and K
post
. In this context I would like to
express my gratitude to Professor Marti Marti from
the technical university in Zrich (ETHZ) and his
scientific assistants, Heinrich Schnetzer and Kyrill
Sokolov, who conducted the first tests with two
different LWAC mixtures, for their outstanding
support. Fig.6 shows stress-strain diagrams of
concretes with different densities at three different
strength levels. All of these stress-strain curves
indicate a stable descending branch as a sign of a
controlled post-failure region. Therefore the diagrams
demonstrates that the described test method presents a
successful way to obtain the full stress-strain curves of
HSLWAC.
Fig.6 also illustrates a very brittle behaviour, also for
LWAC of moderate strength, if aggregates of
moderate density are combined with a high strength
cement matrix. This observation confirms the
definition of HSLWAC given in the introduction.
The different densities in Fig. 6 represent LWAC with lightweight fine
material (ALWAC, lower density), LWAC with natural fine material
(middle density) and normal
concrete (=2,3kg/dm). Typical
for the LWAC with natural fine
material and a strength over 35
MPa was a sliding mode of
failure (Fig.7). In these cases
only the undestroyed regions
outside the damage zone
expanded due to the decreasing
load (moderate snap-back) in
contrast to Fig.3b, where a
horizontal damage zone cant be
defined in that way and where
the extension over nearly the whole specimen length causes an extremely snap-back
effect. Moreover the characteristics of the different mortar-matrix also influence the
properties of the respective concrete with regard to the linearity of the ascending branch,
the ductility in the post-failure region, and finally the failure mode (Fig.8).
Fig.5 Principle of tests using a
linear combination of force and
deformation as control variable
0
5
10
15
20
25
30
35
40
45
0 0,5 1 1,5 2 2,5 3 3,5 4
Strain [mm/m]
C
o
m
p
r
e
s
s
iv
e

s
t
r
e
n
g
t
h

[
M
P
a
]
=1,5 kg/dm
=2,3 kg/dm
cylinder h/=300/100
=1,7 kg/dm
0
5
10
15
20
25
30
35
40
45
50
55
60
0 0,5 1 1,5 2 2,5 3 3,5 4
Strain [mm/m]
C
o
m
p
r
e
s
s
iv
e

s
t
r
e
n
g
t
h

[
M
P
a
]
=2,3 kg/dm
Zylinder h/=300/100
=1,7 kg/dm
=1,9 kg/dm
0
5
10
15
20
25
0 0,5 1 1,5 2 2,5 3 3,5 4
Strain [mm/m]
C
o
m
p
r
e
s
s
iv
e

s
t
r
e
n
g
t
h

[
M
P
a
]
=2,3 kg/dm
cylinder h/=300/100
=1,25 kg/dm
Fig.6 Stress-strain diagrams of concretes with different densities and strength
Fig. 7 Sliding Fig.8 Stress-strain diagram of
mode of failure mortar-matrix with natural and
lightweight fine aggregate
0
5
10
15
20
25
30
35
40
45
0 1 2 3 4
Strain [mm/m]
C
o
m
p
r
e
s
s
i
v
e

s
t
r
e
n
g
t
h

[
M
P
a
]
Normalweight matrix
=2,0 kg/dm
Leightweight matrix
=1,7 kg/dm
4
The European standard ENV 1992-1-4 EC2 [3] recommends a bi-linear stress-strain
diagram (Fig.9) with a constant strain at the break point
(2) and a constant ultimate strain of
lcu
=3.5 . Fig.6
demonstrates that particularly in case of ALWAC the
length of the yield plateau seems to be overestimated in
[3]. In the next months, additional tests will follow to
investigate the behaviour under sustained load. The
purpose of this study is to find a realistic formula which
takes into account the dependence of
lcu
on compressive
strength and oven dry density.
Contrary to the German standard DIN 4219, most of the international standards take into
account that the modulus of elasticity of
LWAC depends on the compressive
strength and the density. The tests
conducted so far with different
aggregates confirm both influences. Fig.
10 shows one evaluation of these results
compared to values calculated by means
of the E-modulus given in EC 2 T.1-1
multiplied with a reduction factor
=(/2,4), where denotes the oven dry
density. Furthermore, investigations will
follow in order to find the reduction
factor with the best possible agreement.
In general, the development of strength
with time is faster in case of LWAC
compared with NC, particularly above the
limit strength of the LWAC (see Fig.14).
Fig.11 shows an evaluation of self tests
and various publications according to the
equation given in Model Code 1990.
Therefore the post-hardening is smaller for
LWAC.
The influence of the specimen-shape on
the compressive strength tend to be
smaller in case of LWAC. Typical values
for the ratio between the compressive strength of cylinders and cubes are slightly above
0,9. The reason for that is the smaller effect of the lateral expansion restraint due to the
smaller lateral strain of LWAC at the maximum load.
0
5
10
15
20
25
30
0 5 10 15 20 25 30
measured E-Modulus [GPa]
c
a
l
c
u
l
a
t
e
d

E
-
M
o
d
u
l
u
s

[
G
P
a
]
=(/2,4)^2
EC 2 T.1-1
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5
Strain
lc
[mm/m]
S
t
r
e
s
s


lc
[
M
P
a
]
f
lcd
=0,8f
lck
/
c
f
lck
Fig.9 Bi-linear stress-strain
diagram given in [3]
Fig.10 Comparison between calculated and
measured values of E-modulus of LWAC
Fig.11 Development of compressive strength
with time for NC and LWAC
5
Fig.12 FE-mesh with geometrical
joints
Fig.13 Relation between concentration of
aggregate volume and LWAC
compressive strength
3 MODELLING OF A LWAC-CUBE UNDER COMPRESSION
An axial compression test of a 100mm
LWAC-cube should be simulated by
means of SBETA, a program for
nonlinear finite element analysis. Several
publications confirmed the strong bond
forming between the matrix and the
expanded aggregates such that failure
would be either through the paste or
through the aggregate. Thus, the concrete
was modelled as a two-component
material for a two-dimensional computer
simulation. In this simulation a plate was
cut from the cube. The laterial strain on
the upper and lower surface of the plate
was eliminated. The load was being
applied on a steel bearing plate controlled
by the strain. The volume of the grains
was chosen with V
p
=40% of the concrete
volume according to self experiences. The
size and the distribution of the 121 pellets
were determined by a statistical
evaluation. Fig.12 shows the FE-mesh
with all geometrical joints. To overcome
the influences of the mesh and three-
dimensional effects, previous analyses
were carried out to calculate the necessary
input value of the compressive strength of
the matrix in a one-component system. In
this way increased matrix strength data
were added to the program to reach the real
value of the axial compression tests with
various matrixes.
At the beginning of the simulation the lack of information about the properties of both
components was indicated in consideration of the stress-strain relation in the compression
and tensile region. In particular, the compressive strength of the aggregates was a
dominate element with regard to the strength of the LWAC-cube. Fig. 13 shows, that in
spite of a high strength matrix the strength of the cube was equal to the strength of the
grains, if the aggregate volume exceeded 40% of the concrete volume. The other way
around, this result can be use to define the compressive strength or efficiency of LWA
respectively.
Fig.14 shows the evaluation of the analysis. The known form demonstrates the
relationship between the compressive strengths of the matrix and the LWAC and
additionally in comparison with laboratory results. The simulation also demonstrates the
different internal stress transfer and the failure mechanisms of LWAC and NC. In general
the aggregates in LWAC have a lower stiffness than the mortar matrix in opposition to
NC. Thus, compressive loads are mainly carried by the stiffer mortar matrix
0
10
20
30
40
50
60
70
80
90
100
0% 5% 10% 15% 20% 25% 30% 35% 40%
Concentration of aggregate volume
C
o
m
p
r
e
s
s
i
v
e

s
t
r
e
n
g
t
h

[
M
P
a
]
Matrix
Lightweight aggregate
Lightweight aggregate concrete
6
Fig.14 Relation between the strengths of Fig.15 Crack propagation in LWAC
matrix and LWAC
corresponding to the stiffness relationship between matrix and aggregates, which causes
transverse stresses in the aggregates and in the matrix. Finally, failure occurs after
exceeding the tensile capacity of the aggregates. The cracks usually propagate straight
through the aggregate particles (Fig. 15). The smooth fracture surfaces transfer less stress
and initiate a brittle failure.
4 PROPERTIES OF THE COMPONENTS
The important meaning of the component properties with regard to the simulation was
already mentioned. Furthermore, their knowledge can be helpful for a better
understanding of the internal stress transfer and for the optimization of the strength. Fig.
16 shows for instance the result of wedge splitting tests with LWACs consisting of
different components. Obviously there seems to exist a strong connection between the
combination of the components and the fracture energy.
natural
sand
50% natural
sand; 50%
lightweight
sand
lightweight
sand
expanded
glass
0
10
20
30
40
50
60
70
80
90
100
F
r
a
c
t
u
r
e

e
n
e
r
g
y

[
J
o
u
l
e
/
m

]
Matrix
P
a
r
t
i
c
l
e

d
e
n
s
i
t
y


p
=1.74 kg/dm

p
=1.22 kg/dm

p
=0.83 kg/dm
0
10
20
30
40
50
60
70
80
90
100
110
0 10 20 30 40 50 60 70 80 90 100 110
Compressive strength of the lightweight matrix[MPa]
M
P
a
]
Aggregate C
Aggregate B
Aggregate A
cube 46
cube 49
cube 38
cube 50
cube 51
cube 42
cube 52
cube 53
cube 54
cube 55
45
Fig. 16 Fracture energy for LWAC with different components,
investigated in wedge splitting tests
7
4.1 Properties of lightweight aggregates
There is a wide range of different lightweight aggregates, which differ in the raw
material, density, shape, outer skin and water absorption. In spite of this fact, their
properties can be estimated with simply formulas, which in general depend on the particle
density. In this context we
tested several aggregates
(expanded clay, shale and
glass, sintered fly ash,
pumice) with regard to the
compressive and tensile
strength. Fig. 17 and 19
demonstrate that there are
obviously connections only
with slight dependence on the
aggregate type. Fig. 17 also
shows the necessity to use a
high-strength matrix in order
to determine the compressive
strength of the particles. The
test methode, measuring the
crushing resistance, can not be
recommend to achieve a
statement about the efficiency of the aggregates.
For the tensile strength of pelletized aggregates, we developed a new testing arrangement
shown in Fig. 18. Twelve grains were glued in the openings of two opposite plates and
the tensile force to the fracture area were related. In this way we achieved the lower limit
of the tensile strength, since a premature failure of a single grain cant be excluded.
Fig.19 shows that the particle tensile strength increases exponential with increasing
particle density. Thus, this fact explains the increasing fracture energy with increasing
particle density.
The dynamic particle modulus of elasticity was estimated according to Schtz as:
E
dyn,p
= 8000
p
2
.
0,0
0,5
1,0
1,5
2,0
2,5
3,0
3,5
4,0
4,5
0,00 0,25 0,50 0,75 1,00 1,25 1,50 1,75 2,00
Particle density
p
[kg/dm]
Argex 550
Argex 650
Argex R
Arlita F3
Arlita F5
Arlita F7
Embra
Leca HD
Leca NW
Liapor 4,25
Liapor 5
Liapor 6
Liapor 6,5
Liapor 7
Liapor 8
Liapor 9,5
Lytag UK
Pollytag
Vasim
Poraver
f
pt
= 1,25
p
1,5
P
a
r
t
i
c
l
e

t
e
n
s
i
l
e

s
t
r
e
n
g
t
h

f
p
t

[
N
/
m
m

]
Fig.18 Testing arrangement for Fig.19 Relation between particle tensile strength
the particle tensile strength and particle density of 7mm grains
0
10
20
30
40
50
60
70
80
90
100
110
0,0 0,2 0,4 0,6 0,8 1,0 1,2 1,4 1,6 1,8 2,0
Particle density
p
[kg/dm]
C
o
m
p
r
e
s
s
i
v
e

c
u
b
e

s
t
r
e
n
g
t
h


[
M
P
a
]
NSM Blhton
LSM Blhton
LSM Blhschiefer
LSM Bims
LSM Blhglas
LSM gesinterte
Steinkohleflugasche
Trendline:
Natural sand matrix
fck,matrix = 82 MPa
f
ck,cube
= 45,6
p
1,5
Trendline:
Lightweight matrix
fck,matrix = 51 MPa
f
ck,cube
= 35,9
p
1,25
Expanded clay
(+ natural sand)
Expanded clay
Expanded shale
Pumice
Expanded glass
Sintered fly ash
Fig.17 Relation between cube compressive strength
of LWAC and particle density in dependence on the
matrix strength
8
4.2 Properties of mortar matrixes
The influences of the different matrixes on the shape of the stress-strain curves were
already mentioned. Additional important aspects are their differences with regard to the
modulus of elasticity and the ratio between compressive strength and tensile strength. For
instance, the internal stress transfer is influenced considerably by the stiffness ratio of
both components. However, the properties of the lightweight matrixes are almost
unknown. This is the reason that a study was started to investigate different matrixes.
Intermediate results will be published at a later moment.
5. CONCLUSIONS
The formulation of general design rules for LWAC requires a deeper understanding of
both components lightweight-aggregate and mortar matrix. There are numerous types of
each component and their combinations, which have a decisive effect on the behaviour of
LWAC. Therefore the investigation of different components was started, which
complement the tests with LWACs in order to
- support the evaluation and explanation of the concrete tests
- optimize the properties of LWACs
- predict the concrete behaviour by means of a computer simulation.
6. ACKNOWLEDGEMENTS
The author would like to thank the sponsor, Lias Franken, for financial support for these
investigations.
7. REFERENCES
[1] Okubo, S., Nishimatsu, Y.: Uniaxial Compression Testing Using a Linear
Combination of Stress and Strain as the Control Variable; Int. J. of Rock Mech.,
Min. Sci. & Geomech., Abstract., Vol. 22, No.5, 1985, pp. 323-330
[2] Jansen, D.C., Shah, S.P.: Effect of Length on Compressive Strain Softening of
Concrete Journal of Engineering Mechanics. Jan., 1997
[3] ENV 1992-1-4 EC2; Design of LWAC with closed structure